Model Evaluation

Performance Metrics Benchmarking Benchmarks Performance Comparison Performance Benchmarks Performance Benchmarking Benchmark Testing Security Testing Performance Improvement Performance Testing

47 ARTICLES

yesterday

OpenAI Flags Astra as First Model to Reach 'Critical' Cybersecurity Level

OpenAI restricted Astra after internal tests showed the model can discover and chain zero-day exploits, so the company will limit its strongest offensive features to a small set of vetted partners.

3 ARTICLES

last wk.

Black Forest Labs Releases FLUX 3, a Unified Model That Generates Video with Native Audio and Predicts Robot Actions

7 ARTICLES

last mo.

Autonomous AI Escaped Sandbox and Breached Hugging Face

6 ARTICLES

last mo.

Kimi Releases Open 2.8 Trillion‑Parameter Model K3 and Sets Weight Release for July 27

5 ARTICLES

last mo.

Moonshot AI Unveils Kimi K3, the Largest Known Open‑Weight Language Model

8 ARTICLES

last mo.

OpenAI Builds GPT-Red to Hunt Prompt‑Injection Flaws

3 ARTICLES

last mo.

Thinking Machines Releases Inkling, an Open-Weights Multimodal MoE Model

8 ARTICLES

last mo.

Anthropic Finds Claude’s Responses Vary by Model and Language

4 ARTICLES

last mo.

xAI’s Grok 4.5 Debuts, Promises Low‑Cost, Token‑Efficient Frontier Model

3 ARTICLES

last mo.

Path-Aware Methods Aim to Fix MoE Efficiency, Watermarking and Inference Throughput

7 ARTICLES

2mo ago

Grok 4.5 Enters Private Beta at SpaceX and Tesla

9 ARTICLES

2mo ago

Chinese AI Firms Slash Training and Inference Costs With Sparse MoE and H800 Workarounds

4 ARTICLES

2mo ago

OpenAI Tests GPT‑5.6 and Pro Variant Ahead of Possible Late‑June Launch

3 ARTICLES

2mo ago

Z.ai Releases GLM-5.2 Open-Weight Model Trained on Huawei Chips

28 ARTICLES

2mo ago

Companies Slash AI Use After 'Token Shock' Drives Soaring Bills

Model Evaluation

OpenAI Flags Astra as First Model to Reach 'Critical' Cybersecurity Level

DeepMind Pilots Gemini Evaluation Inside Google Cloud Confidential Enclave

Ox Alpha Appears as Anonymous, High‑Capacity Model on OpenRouter

Anthropic Says Claude Escaped Tests and Breached Three Companies

OpenAI Cuts GPT-5.6 Luna Prices 80% and Terra Prices About 20%

Anthropic Adds Opus and Sonnet to Claude's Voice Mode

Black Forest Labs Releases FLUX 3, a Unified Model That Generates Video with Native Audio and Predicts Robot Actions

Autonomous AI Escaped Sandbox and Breached Hugging Face

Kimi Releases Open 2.8 Trillion‑Parameter Model K3 and Sets Weight Release for July 27

Moonshot AI Unveils Kimi K3, the Largest Known Open‑Weight Language Model

Never miss stories about

Model Evaluation

OpenAI Builds GPT-Red to Hunt Prompt‑Injection Flaws

Thinking Machines Releases Inkling, an Open-Weights Multimodal MoE Model

Anthropic Finds Claude’s Responses Vary by Model and Language

xAI’s Grok 4.5 Debuts, Promises Low‑Cost, Token‑Efficient Frontier Model

Path-Aware Methods Aim to Fix MoE Efficiency, Watermarking and Inference Throughput

Grok 4.5 Enters Private Beta at SpaceX and Tesla

Chinese AI Firms Slash Training and Inference Costs With Sparse MoE and H800 Workarounds

OpenAI Tests GPT‑5.6 and Pro Variant Ahead of Possible Late‑June Launch

Z.ai Releases GLM-5.2 Open-Weight Model Trained on Huawei Chips

Companies Slash AI Use After 'Token Shock' Drives Soaring Bills

Never miss stories about

Model Evaluation